What is a profitable accounting or IT services firm worth to a buyer who never plans to sell it? One holding company has raised more than $2 billion in new capital to answer that. It owns more than 70 service businesses already. The held detail is the holding period. The model holds firms for good, and it keeps the local name on the door.
The direct answer: Thrive Holdings buys firms in large, fragmented, mission-critical markets. Partner firms keep equity, leadership and local brand. Engineers then work inside each firm and build AI around its real workflows. The pitch to a retiring owner is continuity, not an exit to a flipper.
- The model, in plain terms
- The one efficiency number, with its limit
- Why Canada fits the model now
- What an owner should ask any buyer
The model, in plain terms
The new capital round is over $2 billion, at a $12 billion valuation. Total capital raised since inception is more than $3 billion. The backers include D1 Capital Partners, Altimeter Capital and SoftBank Group. The money scales existing platforms in accounting and IT services. A new platform targets the technical and regulatory work behind essential infrastructure.
The structure differs from a standard roll-up. Firms are not stripped and merged into one brand. Owners keep a stake and keep running their offices. The holding company adds engineers. Those engineers sit with practitioners, learn the workflow, then build and deploy AI inside it. Products built in one firm spread across the platform.
Scale that claim against the map. Canadian AI Companies, the 2026 List Worth Knowing lists the firms building AI in this country. The roll-up model comes from the other direction. It buys firms that never called themselves AI companies, then adds the capability after the purchase.
The one efficiency number, with its limit
The accounting platform reports a clear test. Its tax AI processed 7,000 returns for the 2025 tax year. Preparation time fell by an average of 31%. Data entry accuracy reached up to 98%. Human review stayed in place. Accountants still check the work before a return goes out.
Treat those figures as company-reported. They come from the platform’s own announcements, not an outside audit. Even with that limit, the shape is instructive. The saving lands on preparation, the repeatable middle of the job. The judgement at the end stays human. That split is why the model can scale across 50 or more accounting firms and 2,000 or more professionals without changing who signs the work.
Most AI buyers cannot show even a reported figure. Canadian AI ROI, Only 2% of Businesses Report a Return found returns are rare in Canadian survey data. A buyer who can name returns processed, time saved and review kept gives an owner something testable. Ask for the same three numbers from any buyer at your table.
Why Canada fits the model now
The succession clock is loud in this country. 63% of owners expect to exit within the next decade. The assets involved total more than $2 trillion. Only 6% hold a formal written succession plan. Half hold none at all. Finding a buyer is the most cited barrier, at 47%.
That is the roll-up’s opening. Thousands of profitable service firms have no successor in the family and no buyer in town. AI in Canada, a 2026 Reference for Business Leaders tracks how much of this economy sits in owner-run firms outside the big centres. A permanent buyer with engineers answers two fears at once. The firm survives the owner’s retirement. The staff keep their jobs under the same name.
The model also cuts the other way. AI is handing the second founder the first founder’s bench shows small firms gaining senior-level leverage from AI directly. An owner with energy left can build the same efficiency and keep the whole firm. Selling is one path. It is not the only path the technology opens.
What an owner should ask any buyer
Ask how long the buyer holds its firms, and get the answer in the term sheet. Ask what happens to your name, your staff list and your client contracts in year two. Ask which AI tools arrive after closing, who trains your team on them, and who checks their output. Ask what share you keep, and what that share paid out to owners who sold three years ago.
Then check the answers against your own plan, or the lack of one. Half of owners hold no written plan at all. A buyer’s offer is a bad substitute for a plan, because it prices your firm on the buyer’s timeline. The figure to restate at the end is the one from the survey. 63% of owners expect to exit within ten years. The owners who plan first will meet buyers like this one on their own terms, not on the buyer’s.
